There is a lot of speculation about the growth of PaaS, mostly about how it aligns with the other broad sections of the cloud computing space, namely IaaS and SaaS. The numbers are all over the place. However, based upon our estimations, the PaaS market is expected to reach @ $7.5bn by 2016 and grow at a compound annual growth rate (CAGR) of about 50%.
The bigger story here could be the differential in growth rates of the IaaS, SaaS, and PaaS marketplaces. Or, how they will grow in the next 2 to 5 years.
Figure 1 shows the current market today, in terms of the sizes of the IaaS, PaaS, and SaaS marketplaces. PaaS has always been the smaller portion of the market. Many predicted that PaaS would provide much more relative growth than IaaS and SaaS, and that has not come true.
One of the larger issues is that PaaS and IaaS have really crossed features. There are pure PaaS providers, such as EngineYard. However, the largest PaaS players, including Microsoft and Google, now offer deep IaaS services as well. The 800-pound IaaS public cloud provider, AWS, now offers PaaS capabilities. Thus, it’s hard to determine where the IaaS stops, and the PaaS begins. Or, the other way around.
It’s a fact that IaaS providers will sell more IaaS services with PaaS than without. If you have a platform to build solutions, you’ll build solutions that run on the IaaS portion of the cloud. As a result, I suspect that things being labeled PaaS are not PaaS, and some things that are not labeled PaaS, really are PaaS.
What’s most interesting is what will likely occur next year. As depicted in Figure 2, the growth of IaaS clouds, including Google, AWS, and Microsoft, will likely accelerate. This acceleration will probably occur with both SaaS and PaaS holding steady on their growth rate. However, in proportion, PaaS won’t see the kind of growth most originally estimated, but there will be growth.
Figure 2: In 2015, with a combined market of @ $73 billion, most of the growth will occur in the IaaS and SaaS space, with PaaS only progressing about $2 to $3 billion.
There are a few forces at work here that we need to understand:
- Enterprises are just slow to adopt PaaS. This is largely due to a lack of understanding, and even some of the limitations of public and private PaaS, as they begin to play with this technology. Many of those limitations are being overcome, or have been overcome. However, PaaS continues to be a tough sale, that has not driven enterprises from traditional development tools.
- The specter of lock-in continues to haunt PaaS. While new options such as Docker are beginning to appear, those in enterprise development are afraid that applications built on PaaS clouds won’t be easy to move from cloud-to-cloud. That fear is warranted. Unless you’re will to do some extra work, lock-in will indeed be an issue.
- The PaaS providers themselves have not done a great job of helping define the value of PaaS. While many have good messages, and do educate those within enterprises that are looking at PaaS solutions, PaaS is still confusing to the average IT leader. Many enterprise IT leaders really have no clue where PaaS fits into the development cycle.
- DevOps does not seem PaaS focused. While DevOps is clearly built around the notion of cloud computing, considering the continuous integration, continuous development, and continuous test and deployment, the use of PaaS within a DevOps organization is not well defined, or understood. In many cases, PaaS is left out of the DevOps strategy, or, at least how to leverage that technology properly within a DevOps organization.
- Finally, language and database confusion. Most PaaS providers only support a few programming languages, and a few are proprietary languages. Moreover, native database support is also limited as to what’s offered, with some integration with third party external systems.
While this is not necessarily bad news for private and public PaaS providers, it is a call to action to learn more about the emerging market for this technology, and take some actions. For the larger players, including AWS, Google, and Microsoft, the path might just be to remain on course. Enterprises seem to want to adopt PaaS providers that also provide IaaS, and the other way around.
However, for those PaaS providers that just focus on PaaS, the market problem becomes a bit more complex. Key to selling PaaS will be to make the business and technical case that PaaS provides more value than existing approaches to development. This means finding problem patterns where PaaS is able to knock the ball out of the park, when it comes to providing the right development path.
These problem domains are typically around Dev/Test automation, and other problems where the amount of money required to acquire the development platforms is much reduced when leveraging PaaS. This ultimately means the ability to shift risk from the enterprise to the PaaS provider. This also means that both development and operations problems are solved by the same PaaS technology.
I still find the PaaS space compelling. However, I also understand why there are issues in defining the value for enterprises that are already bombarded with “cloud-everything” messaging these days. The confusion is the enemy of PaaS, and I suspect that this will continue for some time.
I won’t let the PaaS providers off the hook either. They need to do a better job of being thought leaders around the use of PaaS. This should include defining the right and the wrong ways to leverage this technology.
So, do you still bet on PaaS? It’s a path to development, but not the only path to development. You need to be very careful as to where you apply it. Moreover, insist on doing a proof of concept prototype to insure that PaaS is the right choice, and that you’re using the right PaaS technology.